This section is intended to introduce the reader to various aspects of the art that may be related to various aspects of the present invention. The following discussion is intended to provide information to facilitate a better understanding of the present invention. Accordingly, it should be understood that statements in the following discussion are to be read in this light, and not as admissions of prior art.
Transit time ultrasonic flowmeters have exhibited excellent repeatability and absolute accuracy in many flow measurement applications. However, characteristics inherent in the nature of their measurements present difficulties when these meters are applied to custody transfer measurements of petroleum products. A custody transfer takes place when ownership of a batch of a particular product changes. On a small scale, such a transfer takes place at the pump in a gas station, between the owner of the gas station and his customer.
It is industry practice in custody transfer measurements to “prove” the meter; that is, to establish its calibration accurately, by independent means. Provers are usually devices of fixed and precisely established volume. The time required to deliver the volume of product defined by the prover is accurately defined by the transit of a ball or piston, pushed by the product, from one end of the prover to the other. High speed diverter valves initiate the prover run and bypass the prover when the ball reaches the end of its travel. Position switches at the beginning and end of the prover synchronize the proving operation with the operation of the custody transfer meter—the meter to be used to measure the amount of product delivered to a specific customer. The volumetric output measured by the custody transfer meter (in traditional practice, a turbine or positive displacement meter) during the proving run is compared to the volume of the prover and a meter factor (i.e., a calibration correction) is established.
It is also industry practice to perform a set of several prover runs—five is typical—to establish the “repeatability” of the meter factor of the custody transfer meter. Repeatability in the petroleum industry is usually defined as follows: the difference between the high and low meter factors from a set of prover runs, divided by the low meter factor from that set. The repeatability (or in statistical terminology, “range”) of a set of proving runs is a measure of the uncertainty of the meter factor as determined by the average of the results of that set of runs. For example, a repeatability of 0.05% in 5 runs of the prover indicates that the true meter factor for the custody transfer meter lies within a ±0.027% band of the mean meter factor from that run set, with 95% confidence. A meter factor of this accuracy is the accepted standard for custody transfer measurement.
Unlike turbine and positive displacement flow meters, a transit time ultrasonic flow meter does not measure volumetric flow rate continuously, but instead infers it from multiple samples of fluid velocity. Specifically, the volumetric flow rate is determined from periodic measurements of the axial fluid velocity as projected onto one or more acoustic paths—paths along which the transit times of pulses of ultrasound are measured. The path velocity measurements are combined according to rules appropriate to their number and location in the pipe. Many meters employ parallel chordal paths arranged in accordance with a specific method of numerical integration.
The period over which an ultrasonic transit time meter collects a single set of velocity measurements (one velocity measurement or more, depending on the number of paths) is determined by the path transit times, the number of paths, and the data processing capabilities of the meter itself. For liquid meters, the flow samples will typically be collected over periods ranging from 5 to 100 milliseconds, resulting in sample frequencies between 10 Hz and 200 Hz. These figures may differ from one ultrasonic meter design to another.
An ultrasonic flow measurement is thus a sample data system on two counts:                (1) It does not measure the velocity everywhere across the pipe cross section but only along the acoustic paths, and        (2) It does not measure velocity continuously, but instead takes a series of “snapshots” of the velocity from which it determines an average.        
Because of these properties, a transit time ultrasonic meter responds to flow phenomena like turbulence differently than other meters commonly used for custody transfer in the petroleum industry. More specifically, the individual flow measurements of transit time ultrasonic meters will be affected by the small scale random (i.e., turbulent) variations in local fluid velocity. These variations are both temporal and spatial, and an ultrasonic instrument must make multiple measurements to determine the true average flow rate—to reduce the random error contributions due to turbulence to acceptable levels. Turbine meters and positive displacement meters, on the other hand, respond to the flow field in the pipe as a whole; integration of the fluid velocity in space and time is inherent in the nature of their responses. On the other side of the ledger, transit time ultrasonic meters are not encumbered by physical limitations like bypass leakage and friction, and may therefore provide measurement capability over a wider range of velocity and viscosity conditions.
For custody transfer, flow meters are designed to produce pulses per unit volume of fluid that passes through them (for example, 1000 pulses/barrel). The meter factor MF is given by:MF=V/NP
Here                V is the volume of the standard—the prover—between the two position switches embedded in its walls. When a proving run is initiated the flowing fluid is diverted through the prover and pushes a ball or piston past the upstream switch, initiating the run, which is terminated when the ball or piston reaches the downstream switch        NP is the number of pulses produced by the meter during the period which begins when the upstream switch is actuated (time T1) and ends when the downstream switch is actuated (time T2).        
Ultrasonic meters determine a flow rate Q in volume units per second from individual measurements of fluid velocity along one or more acoustic paths. They therefore must generate pulses by means of a frequency converter that produces pulses at a rate k exactly proportional to the volumetric flow rate. Thus the number of pulses NP is given by:NP=kQ(T2−T1)
If the uncertainties in the volume of the standard, the frequency converter k, and the actuations of the upstream and downstream switches are ignored (these terms are generally smaller by an order of magnitude than the uncertainties associated with the flow instrument calibration. In more detailed analyses they are not ignored), the per unit uncertainty in meter factor for a 95% confidence level is given by:dMF/MF=2dQ(N)/Q=2σmean(N)                Where dQ(N) is one standard deviation of the mean of the N flow samples collected during the prove, or σmean(N).        
One standard deviation of the mean, σmean(N), of N representative flow samples taken during a proving run is given by:σmean(N)=S/(N)1/2 
Here, S is the standard deviation of the population of flow samples—the quantitative characterization of the random variability, produced by the turbulence, in the individual flow measurements of the ultrasonic meter, from one flow sample to the next.
An examination of the above equation reveals the variables that must be controlled to achieve satisfactory proving performance in ultrasonic meters: the turbulence intensity as it affects the standard deviation of the flow samples, S, in combination with the number of samples, N, accumulated during each proving run. These parameters must be such that σmean(N) is small enough to ensure that the range of measured meter factors does not exceed the requirement. Calculations indicate that, if σmean(N) can be made small, meters will prove successfully more than 99% of the time.
Meeting these requirements is not straightforward. With a typical line prover operating at nominal flow rate, the duration of a single proving run is about 20 seconds, more or less. If a sample frequency of 50 Hz is assumed, the number of samples that will be collected during a proving run is 20×50=1000. As noted in the previously referenced patent, the random variations due to turbulence in the flow measurements of a four path chordal ultrasonic can be in the 1.75% range (one standard deviation or S) though upstream piping can lead to variations as low as 1.2% or as high as 3%. Substituting the 1.2% figure, 20 second proving runs will produce a σmean(N) of about 0.04%. With the this value of σmean(N), the probability of obtaining a set of 5 proving runs within a 0.05% range is less than 40%, a figure essentially consistent with actual proving experience. Experience also confirms what calculations show: Higher turbulence will produce still smaller probabilities of success.
This, then, is the problem. Turbulence, such as normally encountered in petroleum product pipelines, adversely affects the repeatability of the meter factors for transit time ultrasonic flowmeters, as measured in short duration prover runs. Unless something is done to alter the character of the turbulence, it appears that ultrasonic flowmeter meter factors measured with conventional provers will not achieve repeatability figures meeting petroleum industry expectations.
U.S. Pat. No. 6,647,806 is based on the hypothesis put forward by Dryden. (Hugh L. Dryden and G. B. Schubauer, The Use of Damping Screens for the Reduction of Wind Tunnel Turbulence, Journal of Aeronautical Science, April 1947.) He tied the reduction in turbulence produced by a series of one or more fine mesh screens in cascade to the production of eddies of very small diameters whose energy was dissipated as heat in a settling chamber downstream of the screen(s). Because screens are structurally impractical for resisting the hydraulic forces produced by liquid flow, the means proposed by the patent endeavored to produce the same effect with relatively small holes in plates. It will be seen in the data of Table 1 of that patent, reproduced below, that the improvements achieved were small. The largest reductions in turbulent variations cited in the prior patent were produced by reducers, either alone, or in combination with plates having small holes.
TABLE 1Reproduced from U.S. Pat. No. 6,647,806 B1Standard Deviationof One FlowTurbulence conditioner ConfigurationSampleStraight pipe with no diffuser mechanism1.2% to 1.75%*Large hole perforated plate1.61%Small hole perforated plate0.93%Reducer immediately upstream0.63%Reducer/large hole perforated plate0.64%Reducer/small hole perforated plate0.59%*The lower figure was not included in the referenced patent, but reflects multiple measurements made subsequent to the filing of that patent. Standard deviations higher than 1.75% can be found 5 to 10 diameters downstream of hydraulic disturbances such as bends, compound bends, and header exits.
The method for reducing the effects of turbulence employed by the turbulence conditioners of this invention does not rely on the elimination of turbulence through the dissipation of very small eddies. Rather, the reduction in the random deviations of flow samples is brought about reducing the eddy sizes such that they are effectively averaged within the acoustic beams of the ultrasonic meter.
The reduction in eddy sizes produced by the turbulence conditioners of this invention also leads to an increase in the frequencies of the random variations in fluid velocity produced by the turbulence. The frequency increases also lead to improved proving performance, by making a limited sample of N velocity measurements collected during a proving run more representative of the entire population of velocity variations.